git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: git-cvsimport does not like detached HEAD

From
MLMartin Langhoff <martin@catalyst.net.nz>
Date
Apr 4, 2007, 21:40 UTC
Message-ID
<46141B42.4080102@catalyst.net.nz>
In-Reply-To
<7vmz1pzx12.fsf@assigned-by-dhcp.cox.net>
Junio C Hamano wrote:
Show 7 quoted lines
> In my day-job project, I use git-cvsimport with -i option to
> treat the central CVS repository as if it is just a branch in a
> remote git repository.  Since I hoard many changes in my git
> repository and trickle only a few changes at a time to the
> central CVS back, I keep shuffling and rebasing my commits on
> top of what I obtain from CVS, and many times, my HEAD is
> detached and not on a particular branch.

I don't quite understand detached HEAD very well so my reply may be bogus... (I'm familiar with cg-seek, which seems to be a similar concept of checking out an arbitrary commit in a mode where you are not allowed to commit because you are not really at the tip of any branch).

> However, it seems git-cvsimport does not like to work when your
> HEAD is detached.
Yep - I can see how...
> Here is a minimum patch that seems to let me going, but I am not
> quite sure what the implication is for letting orig_branch (and
> last_branch) to default to "master" (I do not use "master"
> branch for that project so maybe I am getting lucky).

cvsimport should never write to 'master'*, so setting $last_branch to master will mean that it will always switch to the branch it's importing (and the switch will do the initialisation of temp index files, etc). You could set it to '' and get the same effect, perhaps even safer.

* unless the cvs repo has a branch called 'master', and in that case
we're fsck'd.

Which leads me to think -- how magic is master anyway. Should we protect it? The CVS repo could have a branchname collision with us on any refname -- always make sure that repos where I do cvsimport are never development repositories, which means that my dev branches are protected from collisions. I think it's the only sane thing to do.

OTOH, it's trivial to say something like (warning, MUA-broken patch)
diff --git a/git-cvsimport.perl b/git-cvsimport.perl
index 1a1ba7b..73e0544 100755
--- a/git-cvsimport.perl
+++ b/git-cvsimport.perl
@@ -685,6 +685,9 @@ my (@old,@new,@skipped,%ignorebranch);
 # commits that cvsps cannot place anywhere...
 $ignorebranch{'#CVSPS_NO_BRANCH'} = 1;

+# protect 'master';
+$ignorebranch{'master'} = 1;
+
 sub commit {
        if ($branch eq $opt_o && !$index{branch} &&
!get_headref($branch, $git_dir)) {
            # looks like an initial commit


cheers,


m
-- 
-----------------------------------------------------------------------
Martin @ Catalyst .Net .NZ  Ltd, PO Box 11-053, Manners St,  Wellington
WEB: http://catalyst.net.nz/           PHYS: Level 2, 150-154 Willis St
OFFICE: +64(4)916-7224  UK: 0845 868 5733 ext 7224  MOB: +64(21)364-017
      Make things as simple as possible, but no simpler - Einstein
-----------------------------------------------------------------------
Previous: Junio C Hamano
Message 2 of 2 in “git-cvsimport does not like detached HEAD”
  1. Junio C HamanoApr 3, 2007
  2. Martin LanghoffApr 4, 2007

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.